First two-way adaptive brain-computer interface enhances communication efficiency

A team of bioengineers at Tsinghua University, working with medical research colleagues from Tianjin University, both in China, have developed what they describe as the world’s first two-way adaptive brain–computer interface (BCI). In their study published in the journal Nature Electronics, the group used a memristor-based adaptive neuromorphic decoder to build their BCI.

Over the past several decades, bioengineers have developed a variety of BCI devices; some that attach to the scalp, others that work via embedded brain electrodes. What they all have in common is that they listen for brain waves, learning to recognize patterns that can be associated with known thoughts and then listening for those same patterns to carry out a desired behavior—moving a cursor on a screen to a button and pushing it, for example.

In this new study, the team in China brought a whole new dimension to BCI devices by adding technology that allows for feedback directly to the brain, making it a two-way communications device. The whole point of making BCI devices two-way, the team notes, is to improve efficiency and to allow for their use in a wider array of applications. They claim their new device boosts efficiency 100-fold and reduces energy demand by approximately 1,000 times compared to conventional BCI devices.

The new system came about as the research team discovered that brain signal changes are due to interactions with a traditional device. That gave them the idea to create a dual-loop feedback device using a memristor chip—this was chosen due to its neural network architecture and energy efficiency.

The first loop is based on machine learning. It updates the brain wave decoder, allowing it to adapt to changes in signals. The second loop helps the user refine their thoughts to improve control via feedback.

Adding feedback, the researchers note, allows the device to recognize more brain wave patterns, which gives the user the ability to perform more complex tasks. For example, when used with hands-free drone control, the BCI allows additional degrees of freedom, such as rotation and forward-backward motion—all governed exclusively by brain signals.

The researchers suggest their device marks the next step toward the development of a BCI that could eventually allow people with brain damage to regain lost abilities.

© 2025 Science X Network